Howdy folks!
I've got a question about the current levels in my 7591 PP amp.
IIRC, max current rating for a 7591 is 85ma. This may seem silly, but is that for one tube or is that for a pair in PP?
The amp I built is adjustible fixed bias with AC/DC balance pots and test points. Each 7591 has it's own 27ohm resistor to ground at each of the four test points.
If I'm doing this right, then: 27/.9=30ma. Schematic lists 33ma at this bias point so I'm wondering if my math or the schematic is incorrect.
Next question: I know biasing the tube further into class A territory will give better sound at the expense of tube life.
Where exactly is the golden mean here?
I can go: 27/.8=33.75ma
Or: 27/.7=38.6ma
Or: 27/.6=45ma
My HV secondary is rated for 750ma so the extra current is no problem. All HV filter caps are rated for 900v, so voltage-as long as it doesn't exceed max plate and screen ratings-isn't an issue either. If my math above is correct, which of these bias points would give the best balance between sound while sacrificing just a bit more tube life than the most conservative settings?
Thanks for any and all help!
Best,
mr mojo
I've got a question about the current levels in my 7591 PP amp.
IIRC, max current rating for a 7591 is 85ma. This may seem silly, but is that for one tube or is that for a pair in PP?
The amp I built is adjustible fixed bias with AC/DC balance pots and test points. Each 7591 has it's own 27ohm resistor to ground at each of the four test points.
If I'm doing this right, then: 27/.9=30ma. Schematic lists 33ma at this bias point so I'm wondering if my math or the schematic is incorrect.
Next question: I know biasing the tube further into class A territory will give better sound at the expense of tube life.
Where exactly is the golden mean here?
I can go: 27/.8=33.75ma
Or: 27/.7=38.6ma
Or: 27/.6=45ma
My HV secondary is rated for 750ma so the extra current is no problem. All HV filter caps are rated for 900v, so voltage-as long as it doesn't exceed max plate and screen ratings-isn't an issue either. If my math above is correct, which of these bias points would give the best balance between sound while sacrificing just a bit more tube life than the most conservative settings?
Thanks for any and all help!
Best,
mr mojo
hello there mr.mojo
if R = 27 ohms, and you read 0.9V at one of the 7591s cathode, then
V=IR, I=V/R
then it should read I = 0.9/27 = 33.33mA per 7591 tube
again, if you read a 0.6V at one of the cathodes, then your current would be:
I = 0.6/27 = 22.22mA
I should also remind you that this amount of current is actually plate + screen current.
cheers,
JojoD
if R = 27 ohms, and you read 0.9V at one of the 7591s cathode, then
V=IR, I=V/R
then it should read I = 0.9/27 = 33.33mA per 7591 tube
again, if you read a 0.6V at one of the cathodes, then your current would be:
I = 0.6/27 = 22.22mA
I should also remind you that this amount of current is actually plate + screen current.
cheers,
JojoD
Hey Jojo!
Boy do I feel dumb!! I appreciate the clarification. That's what I get by going from memory rather than double checking Ohm's law to be sure!
So, with the math corrected:
1/27=37ma
1.25/27=46ma
Any idea if the 85ma max rating of the 7591 is for a single tube or a pair?
If the 85ma max rating is for a single tube, which of the above bias points would give the best compromise between best sound and moderate tube wear?
Thanks again for your help-I sure do appreciate it!
Best,
mr mojo
Boy do I feel dumb!! I appreciate the clarification. That's what I get by going from memory rather than double checking Ohm's law to be sure!
So, with the math corrected:
1/27=37ma
1.25/27=46ma
Any idea if the 85ma max rating of the 7591 is for a single tube or a pair?
If the 85ma max rating is for a single tube, which of the above bias points would give the best compromise between best sound and moderate tube wear?
Thanks again for your help-I sure do appreciate it!
Best,
mr mojo
Jojo,
Thanks again for the help! Duncans tube spec page isn't working and I couldn't find definative info on whether that spec was for one or two tubes.
I'm running plate voltage at 465v and 33ma. the screens are fed by a pair of 0A3 VR tubes to give a somewhat constant 75v drop.
Max plate for a 7591 is 550v. Max screen is 440v.
If I bias them for 40ma or even a bit more my plate voltage should actually decrease a bit-right?
What I'm looking to do is bias them into relatively high class AB1-not into class A; trying to get the best possible sound, even if it shortens tube life by a year or two. But I don't want to cross the line and shorten tube life by 4 or 5 years or risk outright damage.
The output transformers are rated for 70watts and with AC and DC balance I should in theory be able to bias the 7591s hotter without running the risk of unbalancing the output transformer-is that right?
Or, is this whole situation more complicated than I think it is?
Thanks a lot for the help!
Best,
mr mojo
Thanks again for the help! Duncans tube spec page isn't working and I couldn't find definative info on whether that spec was for one or two tubes.
I'm running plate voltage at 465v and 33ma. the screens are fed by a pair of 0A3 VR tubes to give a somewhat constant 75v drop.
Max plate for a 7591 is 550v. Max screen is 440v.
If I bias them for 40ma or even a bit more my plate voltage should actually decrease a bit-right?
What I'm looking to do is bias them into relatively high class AB1-not into class A; trying to get the best possible sound, even if it shortens tube life by a year or two. But I don't want to cross the line and shorten tube life by 4 or 5 years or risk outright damage.
The output transformers are rated for 70watts and with AC and DC balance I should in theory be able to bias the 7591s hotter without running the risk of unbalancing the output transformer-is that right?
Or, is this whole situation more complicated than I think it is?
Thanks a lot for the help!
Best,
mr mojo
mr mojo said:
Hey Mr.Mojo, I think you haven't understood a very important thing in tubes and power stages.
Max current doesn't say a thing when you must bias an amp: 85mA (yes, it's for each tube) means that if you try to pass more than 85mA the tube will have a very short life.
But when biasing, it's not the current limit that counts, it's the POWER limit. A tube's plate has a finite capacity to dissipate heat, the capacity is indicated in the datasheet as Maximum Plate Dissipation. For those tubes, it's 19W. You must bias the amp mantaining the quiescent plate dissipation under that limit.
Plate dissipation is calculated by multiplying cathode current (I'm semplifying the things a bit) for the plate-to-cathode voltage. So You have to measure plate voltage, write it down on a piece of paper and figure out the max current you can have without exceeding that limit.
So I = P / V, you must divide the Max Plate Dissipation (19W) for the plate voltage to get the maximum current. For 450V, it's 47mA for each tube, MUCH LESS THAN THE 85mA you're taking as "limit".
But... wait a minute!!!! You can't bias the tube at the maximum dissipation in a class AB amp. You cannot bias them at more than 70% of the maximum plate dissipation, so I = 0.7*P / V = (still taking 450V as a guess, you must substitute your plate voltage here) = 33mA for each tube, 66mA for the pair.
Obviously, biasing them to the 70% max Pd will shorten their life but you get good sound, less than 70% you'll increase tube life.
Those considerations are valid in a class AB amp, in a class A amp you can run the tubes to 100% max Pd.
You still can surpass those limits, but the reliability of the amp will be greatly compromised. Some tubes can't stand their maximum Pd (especially cheap chinese ones), imagine what they would do if you surpass that value.
Hope this helps
And for a complete treating of the argument, see here:
http://www.aikenamps.com/Why70percent.html
http://www.aikenamps.com/Biasing.html
http://www.aikenamps.com/Why70percent.html
http://www.aikenamps.com/Biasing.html
Giaime,
WOW! Thank you for taking the time to post such a thoughtful and detailed response-I REALLY appreciate it!
Your response and the links have made things much more clear.
I now know that the max PD safely allowed per tube is (19).7=13.3 watts.
The article from which I built the amp lists 475v on the plates and 33ma cathode current at idle, giving us:
(475).033=15.675 watts; well above the 70% PD!
Since I've got 465 on the plates; staying within the 70% PD rating gives us:
13.3/465=.028ma
And with the 27ohm cathode resistors at the DC balance test points, that gives us:
.028(27)=.756v
So, for my amp the bias needs to be adjusted for .756v at the cathode test points to give me 13.3 watts PD.
What is unclear though, is with a max PD of 13.3 watts, that gives 26.6 wpc as the max safe output from a class AB 7591 amp.
The article states 35 wpc, and at the plate voltage and idle current listed on the schematic gives 31.35 wpc. IIRC, the Scott 299C is said to be rated for 35wpc as well. Are these instantaneous peaks rather than average output?
Once again, thanks so much for your help. I keep this up and maybe in 10 or 15 more years I'll actually start getting a handle on this whole "DIY tube amp hobby!"
Best,
mr mojo
WOW! Thank you for taking the time to post such a thoughtful and detailed response-I REALLY appreciate it!
Your response and the links have made things much more clear.
I now know that the max PD safely allowed per tube is (19).7=13.3 watts.
The article from which I built the amp lists 475v on the plates and 33ma cathode current at idle, giving us:
(475).033=15.675 watts; well above the 70% PD!
Since I've got 465 on the plates; staying within the 70% PD rating gives us:
13.3/465=.028ma
And with the 27ohm cathode resistors at the DC balance test points, that gives us:
.028(27)=.756v
So, for my amp the bias needs to be adjusted for .756v at the cathode test points to give me 13.3 watts PD.
What is unclear though, is with a max PD of 13.3 watts, that gives 26.6 wpc as the max safe output from a class AB 7591 amp.
The article states 35 wpc, and at the plate voltage and idle current listed on the schematic gives 31.35 wpc. IIRC, the Scott 299C is said to be rated for 35wpc as well. Are these instantaneous peaks rather than average output?
Once again, thanks so much for your help. I keep this up and maybe in 10 or 15 more years I'll actually start getting a handle on this whole "DIY tube amp hobby!"
Best,
mr mojo
Hello MrMojo, glad to have been helpful.
Don't confuse plate dissipation with power output, they're two very different things. This leads to another problem...
You see that the amp designer said 33mA, why now we are biasing at 28mA? Don't you sense that something is wrong?
Yes, something's wrong here. That's because we've done some semplifications:
in a pentode, cathode current is equal to plate current PLUS screen current. So what you measure at the cathode is the sum of these two currents. We cannot bias the tube according to the cathode current, because that's not a clear indication on how much current the plate does pass. So, we have to determine REAL plate current. To determine this, simply determine quiescent screen current and subtract this from the cathode current to get plate current.
How to calculate screen current? Surely your amp has some kind of screen resistor connected to the screens of the power tubes (pin 4 and 8), usually one for each tube or one for each pair of tubes. Measure the voltage drop across this resistor: put your meter on the highest voltage scale, and place (carefully, there's high voltage juice here!) the probes across the resistor. You'll get a voltage, now to calculate current you have to do I = V/R, where R is the screen resistor. You should get about 10mA per tube.
Now that we know screen current, we can subtract this to our bias calculations. So now, you have to determine plate current.
We said you must stay under the 70% of the Max Pd, so 0.7 * 19 / 465 = 28mA. But cathode current will be 28mA + the screen current, so to calculate the cathode current you need you must add 28mA to the screen current that you measured. I expect around 38mA here, so you can safely bias your tubes up to 37/38mA. See now, your 33mA value wasn't excessive, in fact that's a little bit conservative, and I must add that you should follow this 33mA rule, because that's not good to use the tube right at its limits.
About the power output, my datasheets for the 7591 says that under class AB1, 450V on the plates, 400V on the screens, 33mA of plate current, 9.4mA of screen current and a 6600 ohm load, you'll get about 45W!
mr mojo said:
Don't confuse plate dissipation with power output, they're two very different things. This leads to another problem...
mr mojo said:
You see that the amp designer said 33mA, why now we are biasing at 28mA? Don't you sense that something is wrong?
Yes, something's wrong here. That's because we've done some semplifications:
in a pentode, cathode current is equal to plate current PLUS screen current. So what you measure at the cathode is the sum of these two currents. We cannot bias the tube according to the cathode current, because that's not a clear indication on how much current the plate does pass. So, we have to determine REAL plate current. To determine this, simply determine quiescent screen current and subtract this from the cathode current to get plate current.
How to calculate screen current? Surely your amp has some kind of screen resistor connected to the screens of the power tubes (pin 4 and 8), usually one for each tube or one for each pair of tubes. Measure the voltage drop across this resistor: put your meter on the highest voltage scale, and place (carefully, there's high voltage juice here!) the probes across the resistor. You'll get a voltage, now to calculate current you have to do I = V/R, where R is the screen resistor. You should get about 10mA per tube.
Now that we know screen current, we can subtract this to our bias calculations. So now, you have to determine plate current.
We said you must stay under the 70% of the Max Pd, so 0.7 * 19 / 465 = 28mA. But cathode current will be 28mA + the screen current, so to calculate the cathode current you need you must add 28mA to the screen current that you measured. I expect around 38mA here, so you can safely bias your tubes up to 37/38mA. See now, your 33mA value wasn't excessive, in fact that's a little bit conservative, and I must add that you should follow this 33mA rule, because that's not good to use the tube right at its limits.
About the power output, my datasheets for the 7591 says that under class AB1, 450V on the plates, 400V on the screens, 33mA of plate current, 9.4mA of screen current and a 6600 ohm load, you'll get about 45W!
A couple of things to keep in mind....
Whether you are dealing with Class AB1 or Class A1, this
has nothing to do with having a 70% rule that someone made up.....You can operate the plate dissipation werever you want up to 100%....you can even go beyond with certain tubes because the data sheets were overly conservative...this is NOT generaly the rule, but the exception.....
Life expectancy is not all about where you set the DC bias point...
The plate load is crucial as well, since that sets the AC power swing...In some amps, the plate load is way too small and the load-line will spend most of it's time swinging past the MAX plate dissipation, while the DC idle current can be well below the MAX plate dissipation...this amp will have the plates glowing slightly red when it is operating at full steam...but when no signal is present, the plates cool down to the idle dissiaption....
WHen it comes to biasing for AB1 amps....it's boils down to your descresion...Technically Class AB1 has a very large biasing range..
Most Class AB1 amps will not be able to be biased into Class A range...the plate dissipation at that point would cause a melt down...
The "standard" biasing method for AB1 that our Hi-Fi forefathers used back in the Golden-era, was for both efficiency and distortion in mind... You basically start cold and work your way hot till you notice the cross-over notch on the scope just fade..
You will NEVER get the cross-over notch 100% gone in AB1 amp...
SInce you are doing this by EYE on a scope, you will have a subjective issue since 5 techs will not all agree at the EXACT point where the notch is "mostly" gone.... So this is not an exact method...but still a good method since you want the point where "most" of the cross-over is gone. ANy further adjustment is just burining up more plate dissipation and is diminishing returns and not efficient...But wait, if you don't care about efficiency and how often you replace tubes and looking for the cleanest sound, then by all means bias it hotter ... just don't get too crazy about it..know your limitations and what to expect for life expectency without melting the tubes down
Some folks prefer to use thier ears and thier eyes...meaning using the scope and then using the ears for final tweak....
especially in guitar amps, where linearity is not always the goal, but adjusting the bias for the right harmonic tone...
If you want to be extrememly anal and scientific about it, you could use a spectrum analyzer or distortion meter to adjust the bias for specific distortion limit...if you can't dial it in then you toss the tubes for better ones ....
Chris
Whether you are dealing with Class AB1 or Class A1, this
has nothing to do with having a 70% rule that someone made up.....You can operate the plate dissipation werever you want up to 100%....you can even go beyond with certain tubes because the data sheets were overly conservative...this is NOT generaly the rule, but the exception.....
Life expectancy is not all about where you set the DC bias point...
The plate load is crucial as well, since that sets the AC power swing...In some amps, the plate load is way too small and the load-line will spend most of it's time swinging past the MAX plate dissipation, while the DC idle current can be well below the MAX plate dissipation...this amp will have the plates glowing slightly red when it is operating at full steam...but when no signal is present, the plates cool down to the idle dissiaption....
WHen it comes to biasing for AB1 amps....it's boils down to your descresion...Technically Class AB1 has a very large biasing range..
Most Class AB1 amps will not be able to be biased into Class A range...the plate dissipation at that point would cause a melt down...
The "standard" biasing method for AB1 that our Hi-Fi forefathers used back in the Golden-era, was for both efficiency and distortion in mind... You basically start cold and work your way hot till you notice the cross-over notch on the scope just fade..
You will NEVER get the cross-over notch 100% gone in AB1 amp...
SInce you are doing this by EYE on a scope, you will have a subjective issue since 5 techs will not all agree at the EXACT point where the notch is "mostly" gone.... So this is not an exact method...but still a good method since you want the point where "most" of the cross-over is gone. ANy further adjustment is just burining up more plate dissipation and is diminishing returns and not efficient...But wait, if you don't care about efficiency and how often you replace tubes and looking for the cleanest sound, then by all means bias it hotter ... just don't get too crazy about it..know your limitations and what to expect for life expectency without melting the tubes down
Some folks prefer to use thier ears and thier eyes...meaning using the scope and then using the ears for final tweak....
especially in guitar amps, where linearity is not always the goal, but adjusting the bias for the right harmonic tone...
If you want to be extrememly anal and scientific about it, you could use a spectrum analyzer or distortion meter to adjust the bias for specific distortion limit...if you can't dial it in then you toss the tubes for better ones ....
Chris
Dear Chris,
biasing by seeing the crossover distortion is a very bad thing. Aiken explained this very well, go read his article
http://www.aikenamps.com/CrossoverNotchBiasing.html
biasing by seeing the crossover distortion is a very bad thing. Aiken explained this very well, go read his article
http://www.aikenamps.com/CrossoverNotchBiasing.html
All due respect to Aiken... but I have my own thoughts on this...
I did look at his article and his first point of positive grid drive does not apply, since I refered to AB1 .....
His second point is valid, but is common sense, you must check your plate disipation at all times even when using a scope..
The 3rd point about plate voltage and cross-over ...depends on number of things and not soo clean and cut...
The 4th and 5th issue about biasing a Push-Pull Class A amp or Single Ended Class-A amp by way of looking at cross-over is rediculously absurd... instead with Class A, you look for symmetric clipping since there is no cross-over...it does not apply since i was refering to AB1...
If your amp can't bias up properly without over running the plate dissipation then there are some fundamental problems in the output stage design...
Chris
I did look at his article and his first point of positive grid drive does not apply, since I refered to AB1 .....
His second point is valid, but is common sense, you must check your plate disipation at all times even when using a scope..
The 3rd point about plate voltage and cross-over ...depends on number of things and not soo clean and cut...
The 4th and 5th issue about biasing a Push-Pull Class A amp or Single Ended Class-A amp by way of looking at cross-over is rediculously absurd... instead with Class A, you look for symmetric clipping since there is no cross-over...it does not apply since i was refering to AB1...
If your amp can't bias up properly without over running the plate dissipation then there are some fundamental problems in the output stage design...
Chris
mr mojo said:
Yes, your plate voltage will decrease a bit when you increase your plate current.
Yes, you can run them 7591s hotter with destroying your opt tranny.
It is not that complicated, I can relate to your situation because this is exactly what I have but a different tube. Instead of 7591 I used EL34, and I play around it a lot especially on the bias where I learn a lot. I even accidentally manage to run the EL34 with 340V plate and 100mA each when I first turned on my amp.
But we all know that tubes are more forgiving of our mistakes.
Giaime said:
hi giaime,
nice articles for a noob like me! i guess the whole point of the articles is that it is possible plate dissipations ratings can be exceeded that puts the output tubes at risk and that biasing can be selected such that this does not happen.
i guess biasing outputs is a balancing act and depending on our priorities one can chose between a safe amp and a good sounding one.
perhaps there is a happy compromise to be found somewhere in between.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.